Pediatric acute myeloid leukemia (AML) is the second most common type of leukemia in children with high relapse rates due to clonal evolution of leukemic stem cells. Current clinical diagnostic methods are not sensitive enough to track these shifts, emphasizing the need for more reliable techniques. Our study addresses this gap by using extracellular vesicles (EVs) to diagnose AML. EVs carry disease-specific biomarkers and reflect the pathological state of their cells of origin. In this study, we have developed a novel CD82 bead-based flow cytometry assay for detailed surface proteomic analysis of EVs sourced from Hematopoietic Stem and Progenitor Cells (HSPCs). Our findings highlight the presence of specific protein signatures on HSPC-derived EVs, indicating a process of EV formation influenced by the originating cell's physiological status and phenotype. The capability of this assay to precisely identify unique surface proteins on EVs holds promise for tracking the progression of leukemia. CD82 has become the pivotal protein of our study due to its critical role in facilitating cellular interactions within the AML environment and a functional tetraspanin involved in EV biogenesis, making the identification of these proteins particularly relevant for understanding the dynamics of leukemic progression through EV-based diagnostics.

We initially focused on CD82-EVs to identify leukemic-specific protein signatures. It became apparent that a single EV population would be insufficient to address the heterogeneity of leukemia. Thus, we expanded our study to analyze other EV subpopulations such as CD9-EVs, CD81-EVs, and CD82-EVs in the plasma of pediatric AML patient samples, which helped us to uncover unique antigen profiles and provided us with a much broader context of different leukemic protein signatures present on each of these EV subpopulations. By doing so, we identified a distinct panel of proteins that are prognostic markers of AML, including CD13, CD36, CD49d, CD106, CD123, and CD252. We discovered these signatures on AML-derived EVs using only a few microliters of plasma, without any cells, which demonstrates the high sensitivity and robustness of our bead-based platform. The discovery of these novel biomarkers on leukemic EV subpopulations not only has diagnostic implications but also has the potential to optimize personalized therapeutic interventions at earlier disease stages. This has been exemplified with several pre-clinical and clinical-stage immunotherapies targeting AML.

The capability to identify and quantify biomarkers present on EVs using only a small amount of plasma can help to characterize 75 different EV proteins. This method demonstrates both its sensitivity and robustness and offers a new non-invasive approach for tracking the progression of leukemia, assessing the effectiveness of treatment, and potentially predicting therapeutic outcomes. This advancement creates new possibilities for non-invasive leukemia diagnosis, monitoring, and personalized treatment strategies in the field of precision immuno-oncology. These findings lay the foundation for developing targeted therapies by identifying proteins involved in AML pathology. The biomarkers found on AML-derived EVs not only reflect the leukemic cells' physiological condition but also likely reflect insights into cellular mechanisms during disease progression and therapy resistance. By targeting these biomarkers, it might be possible to create more effective therapeutic interventions that address the molecular basis of AML, potentially improving patient outcomes.

In addition, our study emphasizes the necessity of refining cryopreservation techniques for the storage of pediatric AML cell samples. We have highlighted the deficiencies in biobank cryopreservation protocols that significantly impact cell viability, contrasting with our laboratory protocol that demonstrates markedly improved cell viability post-cryopreservation by minimizing DMSO exposure and optimizing freezing conditions. These insights are crucial for enhancing the preservation quality of valuable biological specimens for research purposes.

Furthermore, our research underscores the critical need for sophisticated EV isolation and purification techniques to achieve accurate EV sample analysis. The observed absence of DNA in EV-containing fractions challenges the effectiveness of EV-based approaches for tracking genetic mutations. AML is characterized by its heterogeneity, suggesting that techniques such as single-cell transcriptomics applied to peripheral blood mononuclear cells could provide a more sensitive and effective method for mutation detection. This technique bypasses some of the challenges associated with extracting DNA and RNA from EVs and enables the simultaneous sequencing of transcripts. It also allows for the discovery of new mutational subpopulations via clonal hematopoiesis analysis. Adopting this approach could greatly enhance our grasp of AML's intricacies and contribute to the creation of precisely targeted therapies, offering a more nuanced strategy to tackle the disease's complexity and variability.

The identification of specific biomarkers on EVs derived from AML is a significant advancement in the field of leukemia care. This emphasizes the transformative potential of EVs in redefining the diagnosis, monitoring, and treatment of AML, highlighting their critical role in the evolution of personalized medicine. The precise identification of these protein markers significantly improves the capability to track the progression of AML and gauge the effectiveness of therapeutic interventions. Further studies and clinical trials are necessary to validate these biomarkers in a larger cohort of patients. Such efforts are essential for improving patient outcomes and customizing therapeutic approaches for those afflicted with AML, reinforcing the significance of meticulous biomarker analysis to refine leukemia treatment strategies.